They should put their database into readable english text then use the GFDL on *that* for redistribution. Any machine translation of that is a derivative work, so that's protected via the GFDL as well, by the same logic that translating a .wav into a .mp3 is still protected via the copyright.
Posted Oct 23, 2008 4:24 UTC (Thu) by butlerm (subscriber, #13312)
[Link]
That won't work. The sine qua non of copyright is creativity. A usable machine generated natural language representation of a database of facts would not have this property.